In the design of fuel cells and like electrical energy producing devices involving reactant or product gas undergoing electrochemical reaction (process gas), thermal control is a dominant parameter. The electrochemical reactions in such devices are invariably accompanied by heat generation or heat absorpotion because of entropy changes accompanying the reaction and irreversibilities caused by diffusion and activation overpotentials and ohmic resistance. In the accommodation of thermal control, the art has looked to various techniques, none of which are entirely satisfactory.
The thermal control technique seemingly most desirable takes advantage of the sensible heat of the process gas itself as a vehicle for thermal control. Thus, if removal of heat from the cell is desired, the incoming process gas may be supplied to the cell at a temperature lower than the cell operating temperature such that exiting gas removes heat simply by increase in temperature thereof in passage through the cell. In this technique, one adjusts the process gas flow level above the flow level required for production of preselected measure of electrical energy, such additional process gas serving the heat removal function. Disadvantages attending this practice include undesirable pressure drops based on the increased process gas flow, auxiliary power penalty and loss of electrolyte through vaporization or entrainment. By auxiliary power is meant the power requirements of apparatus accessory to the fuel cell proper, e.g., gas pumps, pressurizing systems and the like. As respects electrolyte loss, all process gas in this gas sensible heat technique is in communication with the cell electrolyte in its passage through the cell and, where substantial additional gas is required for thermal control, a very high electrolyte loss due to saturating of the gas with electrolyte vapor is observed in electrolyte gas resulting in quite high electrolyte loss.
In a second thermal control technique, the art has looked to limiting the temperature gradients inside fuel cells by employment of a bipolar plate having an extended fin disposed outside the cell proper, as shown in U.S. Pat. No. 3,623,913 to Adlhart et al. While this technique provides a somewhat more uniform cell temperature, high gas flow passing directly through the cell can result in high electrolyte loss and increased auxiliary power.
A third thermal control technique relies on the sensible heat of a dielectric liquid. Such sensible-heat liquid approach requires much lower auxiliary power as compared to the gaseous heat transfer medium, but requires a separate heat transfer loop and an electrically isolated manifolding system. To avoid shunt currents betwen stacked cells, dielectric fluids such as fluorocarbon or silicon-based oils have been traditionally used as the heat transfer media. Because the catalyst material may be poisoned severely by even a trace amount of these dielectric fluids, a small leak from the heat transfer loop may be fatal to the cell. Also, the dielectric liquids are flammable and have toxic reaction products.
In a fourth technique for thermal control, the art has relied on the latent heat of liquids. Latent heat liquids (U.S. Pat. Nos. 3,498,844 and 3,507,702 to Sanderson; U.S. Pat. No. 3,761,316 to Stedman; and U.S. Pat. No. 3,969,145 to Grevstad et al.) can provide heat transfer at nearly uniform temperature, although there may be some temperature gradients in the stacking direction if the heat transfer plate is placed between a group of cells. The auxiliary power requirements are expected to be extremely low. Suitable dielectric fluids having boiling points in the range of cell operating temperature can be used, but the disadvantages of the sensible-heat liquid approach apply here also. To overcome these disadvantages, non-dielectric media, such as water, can be used. If water is used, a suitable quality steam can be generated for use in other parts of the plant. External heat exchange also is expected to be efficient because of high heat transfer coefficients. Unfortunately, the use of a non-dielectric liquid necessitates elaborate corrosion protection schemes (U.S. Pat. No. 3,969,145 to Grevstad et al.; U.S. Pat. No. 3,923,546 to Katz et al.; U.S. Pat. No. 3,940,285 to Nickols et al.) and/or the use of an extremely low conductivity liquid. During operation, the conductivity may increase, so means to restore the low conductivity may also be required. If the cooling loop is under pressure, good seals are necessary. If a leak develops during the life of the stack because of pinholes caused by corrosion or deterioration of seals, it could paralyze the entire system. Because of the corrosion protection requirements and intricate manifolding, the cost of the heat transfer subsystem operating on dielectric coolant could be substantial.
In a first U.S. patent application, Ser. No. 923,368, commonly-assigned herewith and field on even date, a fundamentally different approach to thermal control of fuel cells is set forth which provides for supplementing the flow of process gas through an electrochemical cell, in measure required for thermal control by sensible heat of process gas, in manner both avoiding electrolyte loss and pressure drop increase across the cell. In implementing this process gas sensible-heat technique, the invention of such commonly-assigned application introduces, in addition to the customary process gas passage in communication with the cell electrolyte, a process gas passage in the cell which is isolated from the cell electrolyte and in thermal communication with a heat-generating surface of the cell. Such electrolyte-communicative and electrolyte-isolated passages are commonly manifolded to a pressurized supply of process gas. The flow levels in the respective passages are set individually by passage parameters to provide both for desired level electrical energy cell output and desired heat removal.
In a second U.S. patent application, Ser. No. 923,369, also commonly-assigned herewith and filed on even date, electrochemical cell structure is set forth for implementing the thermal control technique of said first U.S. patent application wherein such electrolyte-isolated passages are so arranged as to have gas-confining walls contiguous with the electrode served with process gas by the electrolyte-communicative passages. Integral sheet material is preferably corrugated to define channels opening into the electrode and successive alternate channels closed from the electrode by the sheet material.
Apart from the foregoing thermal control techniques, applicants herein have considered the desirability of so-called "reforming" of hydrocarbon content of process gas. Fuel cell gas streams frequently contain methane and other hydrocarbons. The heat value, and hence electrical energy producing potential of methane is about three to four times greater than that of hydrogen. Since methane itself is relatively electrochemically inactive, it is very desirable to reform methane to form hydrogen and carbon monoxide in accordance with the reaction: CH.sub.4 +H.sub.2 O.fwdarw.3H.sub.2 +CO. The hydrogen and carbon monoxide can then participate in the fuel cell reaction either directly or by further water-gas shift. An incentive for carrying out such reforming reaction in a fuel cell is that the reaction is endothermic and would serve to offset heat generated in fuel cell operation due to inherent irreversibility. Accordingly, internal reforming of fuel can reduce the load on the fuel cell cooling system. Introduction of a reforming catalyst in the path of reactive process gas would serve to realize the foregoing advantages. However, since the reforming reaction is endothermic, it creates cold spots for the electrolyte vapor to condense and, in turn, catalyst activity in promoting reforming would be substantially reduced.